<![CDATA[Signature is used as a proof of one’s ratification. This important purpose cause attempts of falsifying the signature. Falsifying signature is relatively simple, because most verification was done manually only by comparing the form of each signature with the originals. It’s even worse since most verification examiner are tired of the boring verification continuously, that’s why the manual signature verification result is found less satisfying. To overcome the weaknesses of this manual verification, this thesis proposes a computerized verification system forLatinos signature, which feature of signature extracted by Localized Arc Pattern Methods according to model pattern of Indonesian signature, which seeks appearance frequency the model pattern of inputted signature image. This feature achieved than compared to the previous reference signature. This comparison yields the dissimilarity value, which is used to define whether the signature is accepted or rejected through a critical value comparison. The experiment achieved an error rate of 9.696%.]]>
Copyrights © 2006
